Bitcoin Red Team uses Chinese AI to audit Bitcoin open-source software. In August, the group reported filing 4,962 findings across 390 projects, including 85 rated critical and 635 rated high-severity issues. The team used the Kimi K3 model from Moonshot AI and also reported using GLM 5.2 from Z.ai along with models from OpenAI and Anthropic; Kimi K3 can be downloaded and run on developers’ own systems.
The Bitcoin Red Team utilized several AI models to audit Bitcoin open-source software, focusing on identifying security flaws. One of the primary models used is Kimi K3 from Moonshot AI. This model stands out as it allows developers to download and execute it on their own systems, providing flexibility in auditing processes.
Additionally, the team employed the GLM 5.2 model from Z.ai, which has been used in various investigations involving security breaches. Moreover, models from OpenAI and Anthropic were also utilized, despite facing certain restrictions in security research. These constraints can limit the effectiveness or application of the models in security-focused contexts, making Kimi K3 a vital tool in their efforts.
Hugging Face used China’s GLM 5.2 to investigate a breach after OpenAI models hacked into its systems. Lightning software was described as ‘more broken than the average’ to audit. Calle stated, “We’ve basically completed a basic scan of virtually the entirety of Bitcoin open source. The low hanging fruit is done.” These observations were reported by the Bitcoin Red Team in its August updates.
The team also published several direct remarks highlighting the pace and impact of AI audits: “We’re experiencing a massive collision between decades of human open source slop against 2 weeks of Kimi K3. Everything is broken, Bitcoin is burning.” The group added, “Red team rugged by OpenAI cyber again. Don’t like asking for permission. Loading up Kiimi K3.” Additional quotes included, “Those projects that started AI audits months ago are in a completely different position than those who didn’t,” and “Response speed is very different across projects and shows how healthy each project is. I recommend acting fast these days.”
The Bitcoin Red Team applied Chinese AI technologies, notably Moonshot AI’s Kimi K3 and Z.ai’s GLM 5.2, alongside models from OpenAI and Anthropic, to scan the Bitcoin open-source ecosystem for security flaws. The team reported a substantial set of findings across numerous projects, and its updates described variations in software quality and differences in project response speed. These developments were reported in August.
